Learning and Instruction 28 (2013) 12e23

Contents lists available at SciVerse ScienceDirect

Learning and Instruction

journal homepage: www.elsevier.com/locate/learninstruc

School context, achievement motivation, and academic engagement:

A longitudinal study of school engagement using a multidimensional
perspective
Ming-Te Wang a, *, Jacquelynne S. Eccles b
a
University of Pittsburgh, 5940 Wesley W. Posvar Hall, Pittsburgh, PA 15260, USA
b
University of Michigan, USA

a r t i c l e i n f o a b s t r a c t

Article history: This longitudinal study adopts a multidimensional perspective to examine the relationships between
Received 30 August 2012 middle school students perceptions of the school environment (structure support, provision of choice,
Received in revised form teaching for relevance, teacher and peer emotional support), achievement motivation (academic self-
18 April 2013
concept and subjective task value), and school engagement (behavioral, emotional, and cognitive
Accepted 20 April 2013
engagement). Participants were from an ethnically diverse, urban sample of 1157 adolescents. The
ndings indicated that student perceptions of distinct aspects of the school environment contributed
Keywords:
differentially to the three types of school engagement. In addition, these associations were fully or
Adolescence
School environment perceptions
partially mediated by achievement motivation. Specically, student perceptions of the school environ-
School engagement ment inuenced their achievement motivation and in turn inuenced all three types of school
Achievement motivation engagement, although in different ways. Moderation effects of gender, ethnicity, and academic ability
Multidimensional perspective were also discussed.
2013 Elsevier Ltd. All rights reserved.

1. Introduction student engagement in school is inuenced by the degree to which

Active engagement in school is critical to student educational
success (Finn & Rock, 1997; Wang & Holcombe, 2010). Youth must
be actively engaged with their school education in order to acquire
the knowledge and skills required for a successful transition into
postsecondary programs and careers (Wang & Eccles, 2012a,
2012b). School engagement is also a malleable state that can be
shaped by school context, therefore holding tremendous potential
as a locus for interventions (Appleton, Christenson, Kim, & Reschly,
2006; Jimmerson, Campos, & Grief, 2003). Currently, and particu-
larly at the secondary level, increasing student engagement is an
explicit goal of many school reform efforts aimed at addressing
problems of student boredom and alienation, low achievement,
and high dropout rate (Marks, 2000).
In order to promote school engagement, we must rst better
understand the school factors that inuence student engagement.
Self-determination theorists suggest that individuals seek experi-
ences that fulll their fundamental needs and identities through
their interaction with the environment. According to this view,
* Corresponding author. Tel.: 4126246945.
E-mail address: mtwang@pitt.edu (M.-T. Wang).
they perceive that the school context meets their psychological
needs (Connell & Wellborn, 1991; Deci & Ryan, 2000; Krapp, 2005).
Stage-environment t and expectancy-value theorists (Eccles et al.,
1993; Roeser, Eccles, & Sameroff, 1998) further argue that the failure
of schools to meet the psychological needs of adolescents often
leads to declines in academic motivation and interest, which in turn
contributes to decreased school engagement and poor academic
performance as adolescents transition to middle school. Most of the
extant research, however, has failed to capture the dynamic and
interactive nature of engagement. Specically, few empirical
studies have focused on how school characteristics interact with
and inuence various forms of school engagement simultaneously.
Moreover, research has not yet fully explained the actual process
that accounts for the effect of the school environment on student
engagement. It is unclear whether various aspects of the school
environment inuence the behavioral, emotional, and cognitive
engagement differentially and whether the associations between
the school environment and engagement are mediated by more
fundamental motivational beliefs within the student. Therefore,
there is a critical need for research that takes an integrative moti-
vational approach to investigate the contextual and psychological
factors that predict school engagement (Fredricks, Blumenfeld, &
Paris, 2004). Such research could contribute to the knowledge

0959-4752/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.learninstruc.2013.04.002
 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23 13

base that informs effective school practices and to the efforts of
researchers and educators who do the important work of identi-
fying the optimal developmental correlates of school engagement.
In this study we adopt a multidimensional perspective, using a
large-scale sample of ethnically diverse students to investigate the
longitudinal associations among school environment, achievement
motivation, and school engagement during the middle school
years. In particular, we expand on previous research by examining
whether school environment inuences student engagement in
school both directly and indirectly through achievement motiva-
tion and whether these associations differ by student gender,
ethnicity, and academic ability. The study of school engagement as
a multidimensional construct, and as an interaction between the
individual and the school environment, will aid in identifying the
particular school characteristics that foster student engagement
and increase our understanding of the mechanisms through which
they operate.
1.1. Theoretical frameworks for school engagement
Self-determination theory and stage-environment t theory posit
that engagement is manifested in the quality of students in-
teractions with learning activities and academic tasks (Deci & Ryan,
2000; Eccles, 2004; Skinner & Wellborn, 1994). Engagement is thus
conceptualized as consisting of multiple distinguishable features,
including behavior, emotion, and cognition (Fredricks et al., 2004;
Jimmerson et al., 2003). Behavioral engagement refers to the actions
and practices that students direct toward school and learning,
including positive conduct and absence of disruptive behavior, as
well as involvement in learning and academic tasks (Connell, 1990;
Finn, 1989). Emotional engagement represents a students positive
affective reactions to, interest in, and valuing of school activities
(Voelkl, 1997). Cognitive engagement refers to students cognitive
investment in learning, including mental efforts directed toward
learning, use of self-regulated strategies to learn and master con-
cepts, and willingness to exert necessary efforts for comprehension
of complex ideas (Corno & Mansinach, 1983; Zimmerman, 1990).
These three components of school engagement are dynamically
embedded within the individual and provide a rich characteriza-
tion of how students act, feel, and think (Wang & Peck, 2013).
School engagement is optimized when students perceive that
the school context fullls their needs for competence, autonomy,
and relatedness (Connell & Wellborn, 1991; Deci & Ryan, 2000).
Competence refers to the need to experience oneself as effective in
ones interactions with the social environment (Elliot & Dweck,
2005), and a students need for competence is fullled when they
know how to effectively achieve desired outcomes (Skinner &
Belmont, 1993). Autonomy refers to the extent to which an
individual experiences oneself as the source of action. Autonomy is
supported when a student perceives schoolwork as relevant to his
or her interests and goals or when a student experiences choice in
determining his or her own behavior (Assor, Kaplan, & Roth, 2002).
Finally, relatedness refers to the need to experience oneself as
connected to other people (Connell & Wellborn, 1991). Fulllment
of the need for relatedness is likely to occur when teachers and
peers create a caring and supportive environment.
Expectancy-value theory provides a theoretical foundation for a
mediational model that links school characteristics to school
engagement and performance through student motivational beliefs
(i.e., academic self-concept and subjective task values). According
to expectancy-value theory, achievement-related choices such as
school engagement are inuenced psychologically by the in-
dividuals expectation for success and subjective valuing of the
academic work (see Eccles, 2007); students most likely to engage in
school learning place higher value and have greater condence in
their academic abilities than those who do not. Expectancy-value
theory also links individual differences in motivational beliefs to
the experiences that individuals have in school contexts. Teachers
create opportunities for students to engage in a variety of school
activities (Eccles et al., 1993; Eccles & Wigeld, 2002), and these
experiences provide students with information regarding their
competency to succeed, their relatedness to others in that setting,
and their autonomy as learnersdallowing them to realize their
personal and social identities (Connell & Wellborn, 1991). This in-
formation cumulates to inuence the development of self-concepts
of ones ability and subjective task values for the types of activities
to which the student is exposed. These motivational beliefs, in turn,
inuence student engagement in various educational activities
(Simpkins, Davis-Kean, & Eccles, 2006). Drawing on these theo-
retical frameworks, school engagement results from an interaction
of the individual with his/her context and is responsive to both
variations in factors of the school environment and motivational
characteristics (see Fig. 1).
1.2. The link between perception of school environment and school
engagement
Research suggests that the t between adolescents psycholog-
ical needs and their school environment inuences both motivation
and school engagement (Alonso-Tapia & Pardo, 2006; Skinner,
Furrer, Marchand, & Kindermann, 2008; Wigeld, Eccles, Davis-
Kean, Roeser, & Scheifele, 2006). Self-determination theorists and
stage-environment t theorists argue that t is optimized when the
school context provides adequate support for the development and
maintenance of a students sense of competence, autonomy, and
relatedness (Deci & Ryan, 2000; Eccles et al., 1993; Wigeld et al.,

School Factors Individual and Psychological Factors Academic Outcomes

Individual
Characteristics
Gender, Race/Ethnicity,
Academic ability

School Characteristics
1. School structure support School Engagement
Achievement Motivation
2. Provision of choice Behavioral engagement
Academic self-concept
3. Teaching for relevance Emotional engagement
Subjective task value
4. Teacher emotional support Cognitive engagement
5. Peer emotional support

Over Time

Fig. 1. Conceptual model.

14 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23
2006). Here we integrate the academic and social aspects of the
school environment and discuss how these school characteristics
either foster or inhibit the fulllment of the basic psychological
needs of adolescents, which in turn impacts their engagement in
school.
School structure support. The need for competence is met when
school is experienced as being optimally structured (Skinner &
Belmont, 1993). Structure refers to the school experiences being
organized in such a way that a student comes to know what it
means and what it takes to be successful in that context (Skinner &
Belmont, 1993). When teachers are clear about their expectations,
provide consistent responses, and adjust instructional strategies to
the level of the student, they provide structure, which supports
greater behavioral participation in academic tasks, and they foster a
stronger sense of connectedness to school amongst students
(Connell, 1990; Urdan & Midgley, 2003). Indeed, Skinner and
Belmont (1993) found that students whose teachers are clear in
their expectations and provide consistent responses are more
behaviorally and emotionally engaged. Thus, we expect that school
structure support will encourage behavioral engagement and
enhance emotional engagement (Hypothesis 1a).
Teachers can support autonomy by allowing students latitude in
their learning activities and by making connections between school
activities and students personal interests and goals (Skinner &
Belmont, 1993; Skinner et al., 2008). We examine two aspects of
school environment in relation to support of student autonomy:
provision of choice and teaching for relevance to the student.
Provision of choice involves students perceptions that their
teachers provide opportunities for participation in decision making
related to academic tasks and school governance and that they
allow for student input into class discussion. These perceived
characteristics predict greater behavioral engagement with school
tasks and higher school identication (Katz & Assor, 2006; Reeve,
Bolt, & Cai, 1999). Eccles et al. (1993) argued that the lack of
decision-making opportunities for students help to explain de-
clines in both interest and valuing of school during the transition to
middle school. Provision of choice can promote school engage-
ment: giving students opportunities to practice their decision-
making skills, to regulate their behavior, and to experience a
sense of personal satisfaction and responsibility as they exert in-
uence upon their learning environment (Assor et al., 2002).
Therefore, we expect that teacher support for student choice and
input will promote behavioral and emotional engagement (Hy-
pothesis 1b).
Teaching for relevance. Students feel a sense of autonomy when
doing work that, rather than simply fullling school requirements,
relates to their interests and has personal meaning (Roeser et al.,
1998); that is, when students believe that the content of the cur-
riculum and the design of instruction provide opportunities for
self-exploration and when the activities provided are meaningful,
relevant, and related to personal interests and goals (Finn & Voelkl,
1993), they feel a sense of autonomy. Additionally, meaningful
curricula and instruction provide an appropriate level of challenge
and a context that can encourage diverse cognitive strategies (e.g.,
opinion formulation, comprehension, and connection) and pro-
mote cognitive engagement (Helme & Clarke, 2001; Wigeld et al.,
2006). Therefore, we expect that teaching for relevance to the
student will encourage both emotional and cognitive engagement
(Hypothesis 1c).
Teacher emotional support has been positively associated with
different indicators of behavioral engagement, including higher
participation in school activities (Battistich, Soloman, Watson, &
Schaps, 1997) and fewer disruptive behaviors (Patrick, Ryan, &
Kaplan, 2007; Ryan & Patrick, 2001; Wang, Brinkworth, & Eccles,
2013). Similarly, in an emotionally supportive and caring school
environment, students are more willing to open ideas up for dis-
cussion, demonstrate more positive attitudes toward academic
studies, and express feelings of enjoyment as they can freely ex-
press themselves and count on teachers for support with a range of
problems (e.g., Furrer & Skinner, 2003; Shim, Cho, & Wang, 2013).
Thus, we expect that feeling emotionally supported by teachers will
increase student behavioral, emotional, and cognitive engagement
(Hypothesis 1d).
Peer emotional support is also important for school engagement,
especially during adolescence when youth have a great need for
peer relatedness. Several studies have demonstrated that adoles-
cents who have positive interactions with peers are more engaged
both behaviorally and emotionally in school (e.g., Wentzel, 2003).
These associations are most likely due to feeling supported and
cared for by ones peers, fullling the adolescent need for related-
ness and promoting the development of a sense of satisfaction in
school. Although there is little evidence that peer emotional sup-
port is related to cognitive engagement, we are inclined to believe
that when adolescents feel peer acceptance for academic achieve-
ment, they develop both condence and competence in discussing
points of view and critiquing each others work. Thus, we expect
that peer emotional support will be positively associated with
behavioral, emotional, and cognitive engagement (Hypothesis 1e).
1.3. Achievement motivation as a mediator between school
environment and engagement
We draw on expectancy-value theories of motivation (Eccles,
2009) in predicting that students motivational beliefs will
mediate the association between perceived school characteristics
and engagement. We assume that student motivation will be
highest when the academic demands of school are a good t for
student expectations of success, educational values and goals, and
individual developmental needs.
Academic self-concept. Self-concept of academic success refers to
an individuals beliefs regarding their capacity to succeed on spe-
cic academic tasks (Eccles, 2009). There is growing evidence that
positive perceptions of the school environment lead to students
having higher levels of condence in their academic self-concept
(Vollmeyer & Rheinberg, 2000; Wigeld et al., 2006). Teachers
who emphasize task structure and relevance foster both a positive
sense of academic self-concept and positive school behaviors,
whereas teachers who fail to provide clear structure and guidelines
for students and who impose meaningless and uninteresting ac-
tivities are in danger of hindering such positive developmental
outcomes (Assor et al., 2002; Roeser et al., 1998). Similarly, when
students feel that they are cared for and treated with respect by
their teachers and peers, they are more likely to develop greater
condence in their academic abilities (Goodenow, 1993; Murdock &
Miller, 2003). Furthermore, when students perceive classroom
material as related to personal interests or goals, they feel more
condent about mastering that material. In terms of provision of
choice, it is not yet known whether providing students with
decision-making opportunities regarding school governance and
course material inuences academic condence. Finally, students
sense of academic self-concept inuences how they behave, how
they feel, and how they thinkdthe three dimensions of engage-
ment (Eccles, 2009), and indeed, student condence in personal
academic ability has been shown to predict level of school
engagement and task involvement (Eccles & Wigeld, 2002;
Schunk, 1996).
Subjective task value of school learning. Subjective task value is
composed of beliefs regarding how enjoyable a task will be, how
useful a task is for fullling short- and long-term goals, and how
well a task meets personal needs and assists the realization of
 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23
personal identities (Eccles et al., 1993; Eccles & Wigeld, 2002).
Several studies have shown that students perceptions of their
school environment predict the value that students attach to
school. Students who experience emotional support and respect
from teachers and peers are more likely to attach greater value to
learning in school (Roeser et al., 1998). Similarly, when students feel
that their teachers convey reasonable and clear expectations, pro-
vide appropriate instrumental help, and support their autonomy,
they are more likely to value a task and experience positive feelings
toward it (Assor et al., 2002; Grolnick, Ryan, & Deci, 1991).
Furthermore, there is strong evidence that the subjective value a
student places on a task inuences the likelihood of his/her selec-
tion and participation in the task (Deci & Ryan, 2000; Updegraff,
Eccles, Barber, & OBrien, 1996). Research shows that motivation
to do schoolwork that is based on interest or value is associated
with various types of school engagement, including active behav-
ioral participation, interest, and self-regulated learning (Connell,
1990; Elliot, McGregor, & Gable, 1999; Katz & Assor, 2006).
Accordingly, we expect students who place high subjective task
value on academic work at school will report higher levels of all
three types of school engagement.
1.4. Moderation effects of gender, ethnicity, and academic ability
When either achievement patterns or perceived school envi-
ronments are examined by gender, ethnicity, or academic ability,
there are often signicant group differences (Malecki & Demaray,
2003; Meece, Glienke, & Burg, 2006; Oates, 2009; Wang, 2009).
For instance, overall, girls report higher levels of school engage-
ment than boys regardless of what type of engagement is consid-
ered (Johnson, Crosnoe, & Elder, 2001; Martin, 2004). However,
according to extant research, the signicance of the role of group
differences (e.g., gender and ethnicity) as a moderator of school
effects on student engagement is unclear.
Little is known about the role that classroom management or
instructional strategies may play in the development of gender
differences in motivation and school engagement. As students
make the transition to middle school, they perceive school as ori-
ented toward teacher control and a social comparison-based
approach to evaluating student ability (Urdan & Midgley, 2003).
While these perceptions of school characteristics can inhibit the
motivation and engagement of both genders in school, some
studies suggest that girls do not respond as well to competitive
teaching practices (Eccles, 2007). Recent studies show that girls and
boys experience different levels of support from teachers and peers
in school (Rueger, Malecki, & Demaray, 2010); thus, it is likely that
teacher and peer support may have different effects on academic
adjustment for each gender.
There is mixed evidence regarding ethnic differences when
examining the relationship between school factors and academic
outcomes; some studies suggest that African American students are
less likely to engage in school than their European American
counterparts (Downey & Ainsworth-Darnell, 2002), while other
studies report no ethnic differences in the relationship between
school climate and student academic adjustment (Harris, 2006;
Smerdon, 1999; Wang & Huguley, 2012). Still other researchers
have found that the impact of positive relationships with teachers
is stronger for African American students than for European
American students (Downey & Ainsworth-Darnell, 2002; Ferguson
& Mehta, 2004).
The effect of school characteristics on student motivation and
engagement may vary as a function of student academic abilities
(Elliot et al., 1999; Pintrich, 2000). For instance, high performing
students are more likely than low performing students to benet
from an autonomous learning environment, as provision of
15
autonomy can reinforce a sense of competence, autonomy, and
academic condence amongst high performing students while
increasing anxiety and a sense of helplessness amongst low per-
forming students (Patrick, Skinner, & Connell, 1993). Low per-
forming students may therefore require greater structure and
support regarding academic achievement. This support may come
from greater clarication of teacher expectations, consistency and
predictability of response, and the employment of less complex
teaching strategies (Skinner & Belmont, 1993). Yet few studies have
examined the moderation effect of student academic ability when
examining associations between school environment, motivation,
and engagement. Including academic ability as a moderator will
allow us to determine whether the level of student engagement in
school is commensurate with demonstrated academic ability.
Taken together, it is evident that there is much to be gained by
considering individual differences in gender, ethnicity, and aca-
demic ability when examining the relations among school char-
acteristics, motivational beliefs, and school engagement. Greater
understanding of these relations will enable researchers and edu-
cators to promote the specic contextual factors known to enhance
student motivational beliefs and engagement in school among
different types of students.
1.5. Research questions and hypotheses
Drawing on self-determination theory, stage-environment t the-
ory, and expectancy-value theory, we investigate how student per-
ceptions of school characteristics predict school engagement when
demographic characteristics and prior school engagement are
controlled for. The three specic research questions of the study are
to examine whether: (Research Question1) student perceptions of
school environment are related differentially to the three di-
mensions of school engagement (i.e., behavioral, emotional, and
cognitive engagement); (Research Question 2) students motiva-
tional beliefs (academic self-concept and subjective task value)
mediate the association between school environment and school
engagement; and (Research Question 3) the strength of associa-
tions among school environment, motivation, and school engage-
ment differ by gender, ethnicity, or academic ability.
For the rst research question, we hypothesize that each aspect
of school environment predicts the three types of school engage-
ment differentially (Hypothesis 1). For the second research ques-
tion, we hypothesize that student perceptions of the school
environment directly and indirectly predict the level of school
engagement through personal motivation beliefs (Hypothesis 2).
Finally, in the absence of a strong empirical basis for making pre-
dictions about whether the associations between perceived school
environment, motivational beliefs, and school engagement will
function differently for boys or girls, for adolescents from different
ethnic groups, or for adolescents with different academic abilities,
we treat the gender, ethnicity, and academic ability moderation
analyses as exploratory (Hypothesis 3).
2. Method
2.1. Participants
Study participants were part of the Maryland Adolescent
Development in Context Study that was designed to increase un-
derstanding of the inuences of social context on behavioral
choices and developmental trajectories during adolescence. Par-
ticipants were from twenty-three public middle schools in a large,
ethnically diverse county near Washington D.C. of the United
States. This study examines two waves of data: Wave 1(N 1157)
collected in early fall of 7th grade and Wave 2 (N 1039) collected
16 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23
at the end of 8th grade. Of these respondents, approximately 56%
are African American, 32% are European American, and 12% are
either biracial or other ethnic minorities. Approximately 52% of the
students are females. Both samples had a normally distributed
range of socioeconomic levels, with the mean pre-tax family annual
income between $45,000 and $49,999 (range: $5000 to >$75,000)
and 86% of primary caregivers reported being employed.
The subsample was based on parent willingness to participate
and on a stratied sampling procedure designed to capture pro-
portional representation of families from each of twenty-three
middle schools. Wave 2 retained approximately 89% of the orig-
inal sample from Wave 1. The most common source of attrition was
relocation outside of the school district. To examine whether
sample attrition inuenced results, we compared individuals with
complete data or missing data at one wave with individuals with
missing data at two or more waves on all indicators included in the
analyses. None of the thirty-one comparisons were statistically
signicant.
2.2. Procedures
Seventh graders were recruited through letters to their families.
Families interested in participating in the study were asked to sign
and return a consent form. The investigators used a mixture of
questionnaires and school records to collect the data. In this study
we used data from the target student and the primary caregiver,
who was most often the mother. Questionnaires with both students
and their primary caregivers were conducted in the home during
each wave of the study. Questionnaire administrators were pri-
marily women with bachelors degrees and their race was the same
as that of the participating adolescent. The questionnaire took
approximately thirty minutes to complete. Participating students
were compensated $20 at each wave of data collection.
2.3. Measures
2.3.1. School engagement
We adapted existing well-established scales to assess the three
dimensions of student engagement in school (e.g., Finn & Voelkl,
1993; Pintrich, 2000; Skinner & Wellborn, 1994). These scales
have been shown to have strong psychometric properties,
including internal consistency, convergent and divergent validity,
and measurement invariance across gender, ethnicity, and socio-
economic status (see Wang, Willett, & Eccles, 2011). All measure-
ments were scaled appropriately, so that high scores indicate
higher levels of school engagement.
The Behavioral Engagement scale included ve items and
measured the extent to which students follow the school rules and
participate in learning activities in school. Sample items are How
often do you get schoolwork done on time? and How often do you
participate in class discussion actively? Responses were based on a
5-point scale, ranging from 1 (almost never) to 5 (almost always).
The six-item scale of Emotional Engagement assessed students
feelings of acceptance, interest, and enjoyment with school. Sample
items are I nd schoolwork interesting and I feel excited by the
work in school. The item responses for the scale range from 1
(strongly disagree) to 5 (strongly agree).
The ve-item scale of Cognitive Engagement measured the
extent to which students use self-regulated learning strategies such
as self-monitoring and evaluation to help understand learning
materials. Sample items are How often do you make academic
plans for solving problems? and How often do you try to relate
what you are studying to other things you know about? Item re-
sponses of the scale range from 1 (almost never) to 5 (almost
always).
2.3.2. School characteristics
The School Environment Measure was used to create ve latent
constructs to assess student perceptions of the school characteris-
tics: (a) School Structure Support, (b) Provision of Choice, (c)
Teaching for Relevance, (d) Teacher Emotional Support, and (e) Peer
Emotional Support. Items for each scale were adapted from existing
well-validated scales (e.g., Midgley et al., 1998; Roeser et al., 1998;
Wigeld, Eccles, & Pintrich, 1996). Previous research has indicated
that these scales have good psychometric properties, including
internal consistency and predictive and criterion validity (Midgley
et al., 1998). The School Structure Support subscale included four
items and assessed teacher clarity of expectations, consistency and
predictability of response, instrumental support, and adjustment of
teaching strategies (e.g., How often do you know what your
teacher expects of you in school?). The Provision of Choice sub-
scale consisted of four items and assessed student opportunities to
make decisions related to academic tasks and school governance
(e.g., How often do you get to choose your partners for group
work?). The Teaching for Relevance subscale included three items
and assessed the extent to which the curriculum and instruction
provide opportunities for self-exploration and offer activities that
are relevant and related to student personal interests and goals
(e.g., How often do you discuss problems and issues that are
meaningful to you?). The Teacher Emotional Support subscale
consisted of three items and assessed student perceived level of
care and support from teachers (e.g., How often can you depend on
teachers to help you out when you have a personal or social
problem at school?). The Peer Emotional Support subscale
included four items and assessed student perceived level of peer
acceptance and positive relationships with peers (e.g., How often
can you depend on your friends to help you out when you have a
social or personal problem at school?). All items were rated on a
ve-point scale, ranging from 1 (not very often) to 5 (very often).
2.3.3. Achievement motivational beliefs
We used two measures adapted from the expectancy and value
scales developed by Eccles et al. (1993) to assess student percep-
tions of achievement motivation. Both scales have been used in
prior studies and demonstrated excellent convergent and
discriminant validity (Anderman, Midgley, Wigeld, & Eccles, 2001;
Jacobs, Lanza, Osgood, Eccles, & Wigeld, 2002). We created two
latent constructs to represent student motivational beliefs in
school: Academic Self-Concept and Subjective Task Valuing of
School Learning, with each construct being represented by multiple
items. The Academic Self-Concept scale included ve items and
assessed students perceptions of their academic ability to learn or
execute courses of action in order to attain educational perfor-
mances (e.g., How good are you in math?). Items were rated on a
seven-point scale, ranging from 1 (not very good) to 7 (very good).
The Subjective Task Valuing of Learning in School scale included
three items and measured students intrinsic interest in and
attainment valuing of academic achievement (e.g., I go to school
because I enjoy learning in school.). Items were rated on a seven-
point scale in terms of importance, ranging from 1 (not an important
reason) to 7 (very important reason).
2.3.4. Covariates
We controlled for a vector of important covariates related to
student socioeconomic characteristics and prior academic
achievement, motivation, and engagement, including gender
(0 female, 1 male), ethnicity (two dummy variables:
0 European American, 1 African American; 0 European
American, 1 others), socioeconomic status (parent education
and family annual income), student prior school engagement, and
academic achievement (GPA) in the early fall of 7th grade. We
 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23
standardized and added parent education and annual family in-
come to create a composite measure of socioeconomic status.
Student academic grade point average (GPA) was collected from
school records. GPA was an average of student grades in the core
academic subjects (English, math, science, and social sciences).
Letter grades were converted into numerical values (A 5, B 4,
C 3, D 2, Failing 1).
2.4. Data analyses
We used structural equation modeling (SEM) with Mplus 6.0 to
t the hypothesized relations among the study constructs (see
Fig. 1). The nested nature of our data (students nested in twenty-
three schools) was accounted for by tting a multilevel model
with random-effects, which produced correctly adjusted standard
errors in the model estimations. We dealt with the missing data
through full-information maximum likelihood estimation, which
allowed us to include all available data and identied the parameter
values that have the highest probability of producing the sample
data (Baraldi & Enders, 2010).
To address the research questions, we began by tting baseline
models that assessed the direct effects of the ve constructs of
school environment on the three constructs of school engagement.
After establishing these direct relationships, we introduced the two
constructs of achievement motivation (academic self-concept and
subjective task value) into the path model and tested their medi-
ating effect on students perceptions of school environment and
school engagement. In order to test for mediation, we estimated
indirect effects with delta method standard errors to conrm the
mediation effects (MacKinnon, Lockwood, Hoffman, West, & Sheets,
2002). Finally, to examine the moderation effect of gender,
ethnicity, and academic ability, we used a stepwise process for the
multigroup comparisons in SEM recommended by Bollen (1989)
and Kenny (2005) to examine whether the measurement and
structural relations in the nal model varied by gender, ethnicity, or
student prior academic achievement level. We included a series of
increasingly restrictive constraints on the model parameters and
examined whether doing so led to a signicant decrease in the
model t by using chi-square difference tests (Chen, 2007). If the
impositions led to a signicant decrease in overall model t, it
would indicate that there were differences across the groups in the
pattern of associations.
3. Results
All bivariate correlations among latent variables were statisti-
cally signicant and in the expected directions (see Table 1). All
variables had low to moderate correlations (from .10 to .49), sug-
gesting little multicollinearity (Kline, 2005). Conrmatory factor
analysis (CFA) veried that the hypothesized constructs measure
discrete, single latent variables. The standardized loadings ranged
from .38 to .84 and were all statistically signicant at the .05 level.
For the assessment of the measurement model, all the latent vari-
ables, including ve school factors, three school engagement fac-
tors, and two achievement motivation factors, were allowed to
intercorrelate simultaneously. The measurement model was found
to provide adequate t, c2 (445, N 1157) 1116.87, p < .001,
CFI .97, TLI .96, RMSEA .02.
3.1. Direct effects of perceived school environment on school
engagement
We present the standardized path coefcients for the nal tted
model in Fig. 2. The overall model t is good, c2 (352, N 1157)
984.32, p < .001, CFI .95, TLI .94, RMSEA .04. The modication
17
Table 1
Means, standard deviations, and intercorrelations among all latent variables
(N 1157).
Variable 1 2 3 4 5 6 7 8 9 10
1. School structure support 1.00
2. Provision of choice .36 1.00
3. Teaching for relevance .37 .31 1.00
4. Teacher emotional .30 .29 .25 1.00
support
5. Peer emotional support .15 .21 .12 .36 1.00
6. Behavioral engagement .29 .19 .11 .21 .18 1.00
7. Emotional engagement .40 .23 .29 .24 .26 .32 1.00
8. Cognitive engagement .13 .13 .31 .20 .23 .23 .30 1.00
9. Academic self-concept .24 .13 .21 .18 .17 .28 .37 .38 1.00
10. Subjective task value .35 .18 .30 .32 .16 .24 .48 .34 .40 1.00
M 3.57 3.44 3.25 2.73 2.99 2.55 3.62 3.58 5.36 4.31
SD 0.85 0.77 0.98 0.92 0.85 0.52 0.60 0.54 1.01 1.45
Internal consistency (a) .85 .77 .80 .70 .87 .72 .77 .81 .84 .77
Note. All coefcients are signicant (p < .01).
indices suggest no signicant cross loadings. Student perceptions of
School Structure, Provision of Choice, and Teacher and Peer
Emotional Support were positively associated with Behavioral
Engagement (gs .30, .10, .18, and .12 respectively). Students who
indicated positive experiences of School Structure, Provision of
Choice, Teaching for Relevance, and Emotional Support from
Teachers and Peers had higher Emotional Engagement (gs .27, .18,
.16, .21 and .22 respectively). Finally, Teaching for Relevance and
Peer Emotional Support were positively associated with Cognitive
Engagement (gs .21 and .20 respectively). Teacher Emotional
Support was not associated with Cognitive Engagement. Overall,
the model accounts for 35% of the variance in Behavioral Engage-
ment, 46% of the variance in Emotional Engagement, and 32% of the
variance in Cognitive Engagement.
3.2. Mediation effects of motivational beliefs
The mediation analyses examined the extent to which student
Academic Self-Concept and Subjective Task Value of Learning
mediated the relationships between perceived school environment
and school engagement (see Fig. 3). The t of this model is
adequate, c2 (520, N 1157) 1417.30, p < .001, CFI .95, TLI .94,
RMSEA .04. Tables 2 and 3 present the results of direct, indirect,
and total effects in the nal model. The relations between School
Structure and three of the engagement variables (Behavioral,
Emotional, and Cognitive Engagement) were partially mediated by
Academic Self-Concept and Subjective Task Value. Moreover, the
relation between Teaching for Relevance and Emotional Engage-
ment was fully mediated by the motivational variables, while the
relation between Teaching for Relevance and Cognitive Engage-
ment was partially mediated by the motivational variables. The
relations between Teacher Emotional Support and both engage-
ment variables (Behavioral and Emotional Engagement) were
partially mediated by the motivational variables. Finally, the rela-
tion between Peer Emotional Support and Behavioral Engagement
was fully mediated by the motivational variables, while the re-
lations between Peer Emotional Support and Emotional and
Cognitive Engagement were partially mediated by motivational
variables.
The inclusion of the mediating variables adds 7%, 13%, and 10% to
the explained variance in Behavioral, Emotional, and Cognitive
Engagement, respectively. Overall, the variables explain 42%, 59%,
and 42% of the variance in Behavioral, Emotional, and Cognitive
Engagement, respectively. The school environment variables
18 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23

Perceptions of School Environment School Engagement

in 7th Grade Early Fall at the end of 8th Grade

R2 = .35
School structure .30***
Behavioral
engagement
.10*
.27*
**
Provision of choice

.18**
R2 = .46
**
.18 Emotional
Teaching for
.16** engagement
relevance
*
.12
.21* .21***
**
Teacher emotional R2 = .32
support
.22 ***
Cognitive engagement

.20**
Peer emotional
support

Fig. 2. Path model depicting direct effect of student perceptions of school environment on school engagement, controlling for prior school engagement, academic achievement, and
demographic characteristics. Only statistically signicant paths were presented in the model for clarity and all coefcients shown were standardized. Paths describing relations from
controlling variables to outcome variables can be obtained in Table 2. *p < .05. **p < .01. ***p < .001.

account for 55% and 47% of the variance in Academic Self-Concept
and Subjective Task Value, respectively.
3.3. Moderation effect of gender, ethnicity, and academic ability
We conducted three sets of moderation analyses to determine
whether the measurement and structural relations in the nal
model varied by gender, ethnicity, or academic ability. For moder-
ation analyses involving ethnicity, we examined differences across
two groups: European American and African American students.
For moderation analyses involving academic ability, we tested
differences across two ability groups: high GPA and low GPA in the
fall of 7th grade (we dened GPA scores above the median as high
and below the median as low). The measurement part of the model
was rst tested by constraining the factor loadings of the two
models to be equal. The Dc2 difference test showed that model t
decrease is not signicant across gender, ethnicity, or academic
ability [gender: Dc2 (23, N 1157) 34.67, p .13; ethnicity: Dc2
(23, N 1157) 33.24, p .21; ability: Dc2 (23, N 1157) 30.58,
p .24]. The constrained measurement model provides a good

Fig. 3. Path model depicting mediational effect of motivational beliefs between perceptions of school environment and school engagement. Only statistically signicant paths were
presented in the model and all coefcients shown were standardized. Direct effects between school environment variables and engagement outcomes were not shown in the model
for clarity but can be obtained in Table 2. Paths describing relations from controlling variables to mediating and outcome variables were presented in Table 3. *p < .05. **p < .01.
***p < .001.
 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23 19

Table 2
Standardized direct, indirect, and total effects for the nal model from perceived school environment and motivational beliefs to school engagement.

Predictor, mediator, and covariate Behavioral engagement Emotional engagement Cognitive engagement

Direct Indirect Total Direct Indirect Total Direct Indirect Total

Predictor variable
School structure support .17** .13** .30*** .15** .14** .29*** .06 .12** .18**
Provision of choice .07 .03 .10* .11* .04 .15* .05 .04 .09
Teaching for relevance .04 .05 .09 .01 .10** .11* .14* .11** .25***
Teacher emotional support .10* .10* .20*** .12* .08* .20*** .05 .08 .13*
Peer emotional support .04 .08* .12* .11* .09* .20*** .12* .11** .23**

Mediator variable
Academic self-concept .29*** e e .18** e e .38*** e e
Subjective task value .19*** e e .37*** e e .16** e e

Covariate
Male .11* e e .03 e e .14* e e
African American .15** e e .17** e e .07 e e
SES .20*** e e .05 e e .10* e e
Prior academic achievement .38*** e e .11* e e .12* e e
Prior behavioral engagement .36*** e e .18** e e .12** e e
Prior emotional engagement .20*** e e .24*** e e .06 e e
Prior cognitive engagement .13* e e .02 e e .38*** e e

Note: *p < .05, **p < .01, ***p < .001.

model t overall for the three groups respectively. After examining
measurement equivalence, the paths of the two models were
constrained to be equal sequentially.
Moderation analyses by gender and ethnicity revealed no sig-
nicant differences in the structural relations and suggest the
generality of the overall model across females and males as well as
European Americans and African Americans. For multigroup ana-
lyses involving academic ability, results indicated signicant dif-
ferences across groups in the relationships between perceived
school environment and motivation as well as between perceived
school environment and school engagement (see Table 4). Further
analyses based on the full mediation model in Fig. 3 revealed that
the academic ability differences were specic to the relationship
between School Structure and Academic Self-Concept, Dc2 (1)
15.03, p < .001, as well as Provision of Choice and Behavioral
Engagement, Dc2 (1) 5.45, p < .05. School Structure was more
strongly associated with Academic Self-Concept for low achievers
(b .24, p < .001) than for high achievers (b .15, p < .01).
Conversely, for high achievers, increasing Provision of Choice was
associated with greater Behavioral Engagement (b .16, p < .01),
while for low achievers increasing Provision of Choice was associ-
ated with lower Behavioral Engagement (b .09, p < .05).
4. Discussion
This study capitalizes on school engagement as a multidimen-
sional constructdincluding behavioral, emotional, and cognitive
engagementdand examines the inuence of multiple school
characteristics on all three types of engagement simultaneously.
The ndings provide empirical evidence regarding the malleability
of school engagement and demonstrate that student perceptions of
the school environment predict change in patterns of school
engagement. For instance, when students experience school as
optimal in structure, have opportunities to make choices, experi-
ence what they learn as having personal relevance, and feel
emotionally supported in learning by both their teachers and peers,
they are more likely to feel interested and to value learning activ-
ities in school. Finally, our study extends the past research on the
application and integration of self-determination theory, stage-
environment t theory, and expectancy-value theory to illustrate the
processes by which different features of the school environment
inuence student behavioral, emotional, and cognitive engagement
in school.
4.1. Direct effects among school environment, achievement
motivation, and engagement

Table 3
The multifaceted nature of engagement captured in this study
Standardized direct, indirect, and total effects for the nal model from perceived makes it possible to extricate the unique contribution of each
school environment and covariates to motivational beliefs. school factor to the three different types of engagement. As we
Predictor and covariate Academic self-concept Subjective task value predicted, the associations between school characteristics and
school engagement vary across the different types of engagement
Direct Indirect Total Direct Indirect Total

Predictor variable
School structure support .19** e e .29*** e e Table 4
Provision of choice .07 e e .08 e e Moderation results: effects of academic ability on modeled relationships.
Teaching for relevance .12** e e .19** e e
Teacher emotional support .12* e e .16* e e
Constrained path c2 (df) p
(for Dc2)
Peer emotional support .22** e e .12* e e
1. Measurement model 1104.46 (512) ns
Covariate 2. School characteristics / motivational beliefs 1145.32 (534) .01
Male .11* e e .03 e e School structure support / academic self-concept 1156.78 (538) .05
African American .15** e e .06 e e 3. School characteristics / school engagement 1179.45 (550) .05
SES .21*** e e .17** e e Provision of choice / behavioral engagement 1072.31 (554) .05
Prior academic achievement .25*** e e .20*** e e 4. Motivational beliefs / school engagement 1102.90 (581) ns

Note: *p < .05, **p < .01, ***p < .001. Note. Bolded pathways denote differences across math ability groups.
20 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23
(Hypothesis 1). For example, in order to increase behavioral
engagement, modication of school structure or the enhancement
of teacherestudent relations appears to be more effective than
fostering relevant academic tasks. It is critical to keep such dis-
tinctions of associations in mind when studying school engage-
ment, as evidently, not all characteristics of school environment
impact the unique types of school engagement in the same manner.
The use of a multidimensional perspective allows us to gain a better
understanding of which aspects of the school environment are
most salient, and hence potentially most effective, in increasing the
different types of school engagement.
As expected, students who report that their teachers provide
clear expectations and consistent and contingent responses are
more likely to participate in academic tasks and to identify in a
positive way with their school (Hypothesis 1a). Similarly, greater
interest and the use of self-regulatory strategies in learning are
more common when the curriculum include meaningful topics that
reect students personal goals and interests (Hypothesis 1c).
Teaching for relevance also predicts higher academic self-concept
and subjective task value. Contrary to our hypotheses, however,
perceived provision of choice did not predict students subjective
task valuing of learning at school, nor did it predict their academic
self-concept (Hypothesis 1b). There is evidence that the benets of
choice-provision for student motivation are likely to be limited if
the choices involve tasks that are not deemed interesting or rele-
vant to a students personal goals and interests (Assor et al., 2002).
Opportunities for decision making or freedom of action are less
important than the extent to which the decision making and action
opportunities available reect personal goals, interests, and values
(Deci & Ryan, 2000). It is possible that decision-making opportu-
nities do not effectively support student autonomy in the case that
no relevance to personal goals or interests is detected. In order to
make choice and freedom of action motivational, students should
be provided with options to engage in schoolwork that are relevant
to personal goals and interests (Flowerday & Schraw, 2003; Katz &
Assor, 2006). A benecial strategy, for example, would be for
teachers to explicitly illustrate and explain the relevance of tasks to
the personal goals and interests of their students when providing
them with choices.
4.2. Moderation effect of academic ability
It is plausible that some students have not developed the
required academic ability to effectively harness the opportunities
afforded to them when choices are made available. In the present
study, the moderation effect of academic ability provides partial
support for this assumption as provision of choice was found to
positively predict the behavioral engagement of high performing
students, but to negatively predict the behavioral engagement of
lower performing students (Hypothesis 3). This moderating impact
of academic achievement on the link between provision of choice
and behavioral engagement suggests that greater behavioral
engagement is more likely to occur in a school environment that
responds to differing levels of academic ability with the develop-
mentally appropriate provision of autonomy (Deci & Ryan, 1985;
Patrick et al., 1993). A second moderation effect of academic ability
supplements this nding: the positive effect of school structure on
academic self-concept is stronger for low achievers. According to
self-determination theory-based studies (see Deci & Ryan, 2000;
Skinner & Belmont, 1993), teachers who balance structural support
and student autonomy in an effort to maintain an engaging envi-
ronment are more likely to encounter increased learning motiva-
tion in their students. For choice to have benecial effects, it needs
to be tailored to student academic ability. Taken together, choice
can be motivational when the task choice options are of
appropriate difculty and when it comes with adequate structure
support. It is important for teachers who provide choice to match
the complexity of the tasks to student academic abilities and
perceived competence. Teaching is most effective when directed
toward the students zone of proximal development, which is
largely determined by their academic skills.
In regard to support for relatedness, we found that students are
more likely to behaviorally and emotionally engaged in school
when teachers and peers create a caring and socially supportive
environment (Hypotheses 1d and e; Battistich et al., 1997; Wentzel,
2003). The ndings add to growing evidence that student percep-
tions of the nature and quality of the school social environment are
as important as the academic environment (e.g., academic tasks
and instructional practices) in promoting adaptive achievement
motivation and engagement in school (e.g., Turner, Meyer, Midgley,
& Patrick, 2003). If focus is placed solely on academic activity, a
teacher runs the risk of creating a negative social environment
wherein students are less likely to become behaviorally and
emotionally engaged, and are more likely to be apprehensive about
making mistakes. Accordingly, if focus is placed on the social di-
mensions of class only while academic dimensions are neglected,
students are less inclined to become behaviorally and cognitively
engaged in learning.
4.3. Mediation effect of achievement motivation
This study identies the role that motivational beliefs play in
mediating the relationship between school environment and
various dimensions of school engagement, demonstrating that
student perceptions of the school environment directly and
indirectly predict the level of school engagement through per-
sonal motivation beliefs (Hypothesis 2). Consistent with our ex-
pectations, an academically adaptive and socially oriented school
environment that facilitates person-environment t promotes
student engagement through its inuence on student academic
self-concept and subjective task values. In particular, student
academic self-concept and subjective task valuing of learning
is enhanced when the school environment provides clarity
of expectation, consistency and predictability of response,
emotional support, opportunity to learn and master meaningful
material, and sufcient or appropriate support of students per-
sonal goals and interests. This study informs the understanding
of the mechanisms by which the school environment inuences
student engagement, thus highlighting critical points of
intervention.
Understanding student engagement in school requires an inte-
grative motivational framework that considers the interaction of
the psychological and contextual factors in a youths life, as well as
the developmental needs of the youth as he or she matures within
these contexts. Most studies of school engagement and achieve-
ment in motivational psychology focus on a single theoretical
framework rather than taking advantage of the areas rich theo-
retical landscape. The phenomenon of looking for lost keys under
the lamppost rather than in all the other places they might be
epitomizes this focused, but decient approach. Through the
implementation of multiple motivational theories (self-determina-
tion theory, stage-environment t theory, and expectancy-value the-
ory) in the design of this research and in the interpretation of the
results, this study provides a much richer picture of the role of
school engagement in adolescent development.
4.4. Limitations, future research, and implications for practice
Researchers have typically used student self-reports of school
characteristics in order to comprehend how students themselves
 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23
derive meaning from their own perceptions of the school learning
environment. In fact, a growing number of studies have shown that
student perceptions of school experiences are critical components
in understanding their developmental outcomes (e.g., Wang &
Holcombe, 2010). However, relying on student self-reports in
assessing perception of school context raises an important validity
concern, which is that students could answer questions about their
behavior or that of their teacher in ways that they perceive to be
socially desirable. Future inclusion of multiple sources of informa-
tion (teachers, principals, parents), as well as multiple methodol-
ogies (interviews, observations, surveys), will provide a more
robust, valid method of identifying school effects related to
engagement. Furthermore, research from our motivational frame-
works presumes a causal sequence that the perceived school
environment contributes to individual motivational beliefs. How-
ever, it also has been suggested that the extent to which students
are motivated in school may also inuence their responses to the
school environment (Urdan & Midgley, 2003). It is possible that
students with high academic self-concept and particularly high
task value are more likely to attend to and appreciate teachers
reference to relevance. Thus, future research should examine the
prospective reciprocal relations between school environment and
adolescent motivational beliefs. Additionally, in this study we
examined the impact of school context on student engagement in
school; a more comprehensive examination would take into ac-
count other contexts such as family and neighborhood. Finally,
school engagement is likely to take different forms in the
elementary and secondary school years (Fredricks et al., 2004).
Future comparison of longitudinal studies across different age
ranges would be of use in the clarication of whether the ndings
identied in this study were unique to this sample at this given
time, or whether the ndings in fact represent an otherwise
generalizable or even universal developmental trajectory for school
engagement. Greater precision in the articulation of the develop-
mental effects of school engagement will help to inform in-
terventions to support adolescent learning at different levels of
schooling.
The study of school engagement as a multidimensional
construct and as an interaction between the individual and the
school environment enabled us to identify which particular school
characteristics foster or undermine student engagement and to
achieve a greater understanding of the psychological mechanisms
through which the school environment inuences student
engagement in school. The current study suggests that teachers
have a critical role to play in creating a positive school environment,
which can in turn help them to optimize their students perceptions
of their school environment. The ve aspects of school environ-
ment examined in this study are inuenced by the academic goals
and social climate emphasized and fostered by teachers in their
implementation of school policies, selection of instructional prac-
tices and academic tasks, organization of the classroom environ-
ment, and the manner in which they relate to their students.
Teachers can promote student learning motivation and engage-
ment by creating a supportive school environment that stresses
and provides opportunities for students to feel competent, auton-
omous, and emotionally supported (Al, Assor, & Katz, 2004; Reeve,
Jang, Carrell, Jeon, & Barch, 2004). Such school environments
enable all students to succeed in the required academic and social
tasks at hand and to grasp the role that schoolwork has to play in
the realization of their personal goals, interests, and values. The
present study thus supports the notion that attending to the quality
of both academic and social domains of the school environment is
important in developing a positive student perception of school,
which in turn promotes student motivation and engagement. With
a thorough understanding of how school characteristics affect
21
student behavioral, emotional, and cognitive orientation, schools
will be more able to establish practices that prevent disengagement
and to foster learning environments that support engagement e
resulting in the mutually benecial outcome of enhanced academic
performance.
References
Al, O., Assor, A., & Katz, I. (2004). Learning to allow temporary failure: potential
benets, supportive practices and teacher concerns. Journal of Education for
Teaching, 30, 27e41. http://dx.doi.org/10.1080/0260747032000162299.
Alonso-Tapia, J., & Pardo, A. (2006). Assessment of learning environment
motivational quality from the point of view of secondary and high school
learners. Learning and Instruction, 16, 295e309. http://dx.doi.org/10.1016/
j.learninstruc.2006.07.002.
Anderman, E. M., Midgley, C., Wigeld, A., & Eccles, J. S. (2001). Learning to value
mathematics and reading: relations to mastery and performance-oriented
instructional practices. Contemporary Educational Psychology, 26, 76e96.
http://dx.doi.org/10.1006/ceps.1999.1043.
Appleton, J. J., Christenson, S. L., Kim, D., & Reschly, A. L. (2006). Measuring cognitive
and psychological engagement: validation of the student engagement instru-
ment. Journal of School Psychology, 44, 427e445. http://dx.doi.org/10.1016/
j.jsp.2006.04.002.
Assor, A., Kaplan, H., & Roth, G. (2002). Choice is good, but relevance is excellent:
autonomy-enhancing and suppressing teacher behaviors predicting students
engagement in schoolwork. British Journal of Educational Psychology, 72, 261e
278. http://dx.doi.org/10.1348/000709902158883.
Baraldi, A. N., & Enders, C. K. (2010). An introduction to modern missing data
analyses. Journal of School Psychology, 48, 5e37. http://dx.doi.org/10.1016/
j.jsp.2009.10.001.
Battistich, V., Soloman, D., Watson, M., & Schaps, E. (1997). Caring school
communities. Educational Psychologist, 32, 137e151. http://dx.doi.org/10.1207/
s15326985ep3203_1.
Bollen, K. A. (1989). Structural equations with latent variables. Hoboken, N.J.: Wiley.
Chen, F. F. (2007). Sensitivity of goodness of t indexes to lack of measurement
invariance. Structural Equation Modeling, 14, 464e504. http://dx.doi.org/
10.1080/10705510701301834.
Connell, J. P. (1990). Context, self, and action: a motivational analysis of self-system
processes across the life-span. In D. Cicchetti (Ed.), The self in transition: From
infancy to childhood (pp. 61e97). Chicago: University of Chicago Press.
Connell, J. P., & Wellborn, J. G. (1991). Competence, autonomy, and relatedness: a
motivational analysis of self-system processes. In M. R. Gunnar, & L. A. Sroufe
(Eds.). Self processes in development: Minnesota symposium on child psychology,
Vol. 23, (pp. 43e77). Chicago: University of Chicago Press.
Corno, L., & Mansinach, E. (1983). The role of cognitive engagement in classroom
learning and motivation. Educational Psychologist, 18, 88e108. http://dx.doi.org/
10.1080/00461528309529266.
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human
behavior. New York: Plenum Press.
Deci, E., & Ryan, R. M. (2000). What is the self in self-directed learning? Findings
from recent motivational research. In G. Staka (Ed.), Conceptions of self-directed
learning: Theoretical and conceptual considerations). Munster: Waxmann.
Downey, D. B., & Ainsworth-Darnell, J. W. (2002). The search for oppositional cul-
ture among black students. American Sociological Review, 67, 156e164.
Eccles, J. S. (2004). Schools, academic motivation, and stage-environment t. In
R. M. Lerner, & L. Steinberg (Eds.), Handbook of adolescent psychology (2nd ed.).
(pp. 125e153) Hoboken, NJ: Wiley.
Eccles, J. S. (2007). Where are all the women? Gender differences in participation in
physical science and engineering. In S. J. Ceci, & W. M. Williams (Eds.), Why
arent more women in science? Top researchers debate the evidence (pp. 199e210).
Washington, DC: American Psychological Association. http://dx.doi.org/
10.1037/11546-016.
Eccles, J. S. (2009). Who am I and what am I going to do with my life? Personal and
collective identities as motivators of action. Educational Psychologist, 44, 78e89.
http://dx.doi.org/10.1080/00461520902832368.
Eccles, J. S., Midgley, C., Buchanan, C. M., Flanagan, C., Mac Iver, D., Reuman, D.,
et al. (1993). Development during adolescence: the impact of stage/envi-
ronment t. American Psychologist, 48, 90e101. http://dx.doi.org/10.1037/
0003-066X.48.2.90.
Eccles, J. S., & Wigeld, A. (2002). Motivational beliefs, values, and goals.
Annual Review of Psychology, 53, 109e132. http://dx.doi.org/10.1037/0003-
066X.48.2.90.
Elliot, A. J., & Dweck, C. S. (Eds.). (2005). Handbook of competence and motivation).
New York: Guilford.
Elliot, A. J., McGregor, H. A., & Gable, S. (1999). Achievement goals, study strategies,
and exam performance: a meditational analysis. Journal of Educational
Psychology, 91, 549e563. http://dx.doi.org/10.1037/0022-0663.91.3.549.
Ferguson, R. F., & Mehta, J. (2004). An unnished journey: the legacy of Brown
and the narrowing of the achievement gap. Phi Delta Kappan, 85, 656e669.
http://dx.doi.org/10.2307/i20189399.
Finn, J. D. (1989). Withdrawing from school. Review of Educational Research, 59, 117e
142. http://dx.doi.org/10.3102/00346543059002117.
22 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23
Finn, J. D., & Rock, D. A. (1997). Academic success among students at risk for school
failure. Journal of Applied Psychology, 82, 221e234. http://dx.doi.org/10.1037/
0021-9010.82.2.221.
Finn, J. D., & Voelkl, K. E. (1993). School characteristics related to school engage-
ment. Journal of Negro Education, 62, 249e268. http://dx.doi.org/10.2307/
i314505.
Flowerday, T., & Schraw, G. (2003). Effect of choice on cognitive and affective
engagement. Journal of Educational Research, 96, 207e215. http://dx.doi.org/
10.2307/i27542431.
Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). School engagement: potential
of the concept, state of the evidence. Review of Educational Research, 74, 59e
109. http://dx.doi.org/10.3102/00346543074001059.
Furrer, C., & Skinner, E. (2003). Sense of relatedness as a factor in childrens aca-
demic engagement and performance. Journal of Educational Psychology, 95,
148e161. http://dx.doi.org/10.1037/0022-0663.95.1.148.
Goodenow, C. (1993). Classroom belonging among early adolescent student: re-
lationships to motivation and achievement. Journal of Early Adolescence, 13, 21e
43. http://dx.doi.org/10.1177/0272431693013001002.
Grolnick, W. S., Ryan, R. M., & Deci, E. L. (1991). Inner resources for school
achievement: motivational mediators of childrens perceptions of their parents.
Journal of Educational Psychology, 83, 508e517. http://dx.doi.org/10.1037/0022-
0663.83.4.508.
Harris, A. L. (2006). I (dont) hate school: revisiting oppositional culture theory of
blacks resistance to schooling. Social Forces, 86, 797e834. http://dx.doi.org/
10.2307/i405623.
Helme, S., & Clarke, D. (2001). Identifying cognitive engagement in the mathematics
classroom. Mathematics Education Research Journal, 13, 133e153. http://
dx.doi.org/10.1007/BF03217103.
Jacobs, J. E., Lanza, S., Osgood, D. W., Eccles, J. S., & Wigeld, A. (2002). Changes in
childrens self-competence and values: gender and domain differences across
grades one through twelve. Child Development, 73, 509e527. http://dx.doi.org/
10.2307/i370705.
Jimmerson, S. R., Campos, E., & Greif, J. L. (2003). Toward an understanding of
denitions and measures of school engagement and related terms. The Cali-
fornia School Psychologist, 8, 7e27.
Johnson, M. K., Crosnoe, R., & Elder, G. (2001). Students attachment and academic
engagement: the role of the race and ethnicity. Sociology of Education, 74, 318e
340. http://dx.doi.org/10.2307/i326042.
Katz, I., & Assor, A. (2006). When choice motivates and when it does not. Educa-
tional Psychology Review, 19, 429e442. http://dx.doi.org/10.1007/s10648-006-
9027-y.
Kenny, D. A. (2005). Multiple group models. Retrieved 20.04.06, from. http://
davidkenny.net/cm/mgroups.htm.
Kline, R. B. (2005). Principles and practice of structural equation modeling (2nd ed.).
New York, NY: Guilford Press.
Krapp, A. (2005). Basic needs and the development of interest and intrinsic moti-
vational orientations. Learning and Instruction, 15, 381e395. http://dx.doi.org/
10.1016/j.learninstruc.2005.07.007.
MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002).
A comparison of methods to test mediation and other intervening variables
effects. Psychological Methods, 7, 83e104. http://dx.doi.org/10.1037/1082-
989X.7.1.83.
Malecki, C. K., & Demaray, M. K. (2003). What type of support do they need?
Investigating student adjustment as related to emotional, informational,
appraisal, and instrumental support. School Psychology Quarterly, 18, 231e252.
http://dx.doi.org/10.1521/scpq.18.3.231.22576.
Marks, H. M. (2000). Student engagement in instructional activity: patterns in
elementary, middle and high school years. American Educational Research
Journal, 37, 153e184. http://dx.doi.org/10.2307/i248902.
Martin, A. J. (2004). School motivation of boys and girls: differences of degree,
differences of kind, or both? Australian Journal of Psychology, 56, 133e146 http://
dx.doi.org/10.1080/00049530412331283363.
Meece, J. L., Glienke, B. B., & Burg, S. (2006). Gender and motivation. Journal of
School Psychology, 44, 351e373. http://dx.doi.org/10.1016/j.jsp.2006.04.004.
Midgley, C., Kaplan, A., Middleton, M., Maehr, M., Urdan, T., & Hicks Anderson, L.
(1998). The development and validation of scales assessing students achieve-
ment goal orientations. Contemporary Educational Psychology, 23, 113e131.
http://dx.doi.org/10.1016/j.bbr.2011.03.031.
Murdock, T. B., & Miller, A. D. (2003). Teachers as sources of middle school students
motivational identity: variable-centered and person-centered analytic ap-
proaches. Elementary School Journal, 103, 383e399. http://dx.doi.org/10.2307/
i241865.
Oates, G. L. St (2009). An empirical test of ve prominent explanations for the
blackewhite academic performance gap. Social Psychology Education, 12, 415e
441. http://dx.doi.org/10.1007/s11218-009-9091-5.
Patrick, B. C., Skinner, E. A., & Connell, J. P. (1993). What motivates childrens
behavior and emotion? Joint effects of perceived control and autonomy in the
academic domain. Journal of Personality and Social Psychology, 65, 781e791.
http://dx.doi.org/10.1037/0022-3514.65.4.781.
Patrick, H., Ryan, A. M., & Kaplan, A. (2007). Early adolescents perceptions of
the classroom social environment, motivational beliefs, and engagement.
Journal of Educational Psychology, 99, 83e98. http://dx.doi.org/10.1037/0022-
0663.99.1.83.
Pintrich, P. R. (2000). The role of goal orientation in self-regulated learning. In
M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.), Handbook of self-regulation:
Theory, research, and application (pp. 451e502). http://dx.doi.org/10.1016/
B978-012109890-2/50043-3.
Reeve, J., Bolt, E., & Cai, Y. (1999). Autonomy supportive teachers: how they teach
and motivate students. Journal of Educational Psychology, 91, 537e548. http://
dx.doi.org/10.1037/0022-0663.91.3.537.
Reeve, J., Jang, H., Carrell, D., Jeon, S., & Barch, J. (2004). Enhancing students
motivation by increasing teachers autonomy support. Motivation and Emotion,
28, 147e169. http://dx.doi.org/10.1023/B:MOEM.0000032312.95499.6f.
Roeser, R. W., Eccles, J. S., & Sameroff, A. J. (1998). Academic and emotional
functioning in early adolescence: longitudinal relations, patterns, and pre-
diction by experience in middle school. Development and Psychopathology, 10,
321e352.
Rueger, S. Y., Malecki, C. K., & Demaray, M. K. (2010). Relationship between multiple
sources of perceived social support and psychological and academic adjustment
in early adolescence: comparisons across gender. Journal of Youth and Adoles-
cence, 39, 47e61. http://dx.doi.org/10.1007/s10964-008-9368-6.
Ryan, A. M., & Patrick, H. (2001). The classroom social environment and
changes in adolescents motivation and engagement during middle school.
American Educational Research Journal, 28, 437e460. http://dx.doi.org/
10.2307/i360666.
Schunk, D. H. (1996). Goal and self-evaluative inuences during childrens cognitive
skill learning. American Educational Research Journal, 33, 359e382. http://
dx.doi.org/10.2307/i248887.
Shim, S. S., Cho, Y., & Wang, C. (2013). Classroom goal structures, social
achievement goals, and adjustment in middle school. Learning and Instruction,
23, 69e77.
Simpkins, S. D., Davis-Kean, P., & Eccles, J. S. (2006). Math and science moti-
vation: a longitudinal examination of the links between choices and be-
liefs. Developmental Psychology, 42, 70e83. http://dx.doi.org/10.1037/0012-
1649.42.1.70.
Skinner, E. A., & Belmont, M. J. (1993). Motivation in the classroom: reciprocal effect
of teacher behavior and student engagement across the school year. Journal of
Educational Psychology, 85, 571e581. http://dx.doi.org/10.1037/0022-
0663.85.4.571.
Skinner, E., Furrer, C., Marchand, G., & Kindermann, T. (2008). Engagement
and disaffection in the classroom: part of a larger motivational dynamic.
Journal of Educational Psychology, 100, 765e781. http://dx.doi.org/10.1037/
a0012840.
Skinner, E. A., & Wellborn, J. G. (1994). Coping during childhood and adolescence: a
motivational perspective. In D. Featherman, R. Lerner, & M. Perlmutter (Eds.).
Life-span development and behavior, Vol. 12, (pp. 91e133). Hillsdale, NJ:
Erlbaum.
Smerdon, B. A. (1999). Engagement and achievement: differences between African-
American and white high school students. Research in Sociology of Education and
Socialization, 12, 103e134.
Turner, J. E., Meyer, D. K., Midgley, C., & Patrick, H. (2003). Teacher discourse and
sixth graders reported affect and achievement behaviors in two high-mastery/
high performance mathematics classrooms. The Elementary School Journal, 103,
357e382. http://dx.doi.org/10.2307/i241865.
Updegraff, K. A., Eccles, J. S., Barber, B. L., & OBrien, K. M. (1996). Course enrollment
as self-regulatory behavior: who takes optional high school math course.
Learning and Individual Differences, 8, 239e259.
Urdan, T., & Midgley, C. (2003). Changes in the perceived classroom goal structure
and pattern of adaptive learning during early adolescence. Contemporary
Educational Psychology, 28, 524e551. http://dx.doi.org/10.1016/S0361-476X(02)
00060-7.
Voelkl, K. E. (1997). Identication with school. American Journal of Education, 105,
294e319.
Vollmeyer, R., & Rheinberg, F. (2000). Does motivation affect performance via
persistence? Learning and Instruction, 10, 293e309.
Wang, M. T. (2009). School climate support for behavioral and psychological
adjustment: testing the mediating effect of social competence. School Psychol-
ogy Quarterly, 24, 240e251. http://dx.doi.org/10.2307/i245239.
Wang, M. T., Brinkworth, M. E., & Eccles, J. S. (2013). The moderation effect of
teacherestudent relationship on the association between adolescents self-
regulation ability, family conict, and developmental problems. Develop-
mental Psychology, 49, 690e705. http://dx.doi.org/10.1037/a0027916.
Wang, M. T., & Eccles, J. S. (2012a). Adolescent behavioral, emotional, and cognitive
engagement trajectories in school and their differential relations to educational
success. Journal of Research on Adolescence, 22, 31e39. http://dx.doi.org/10.1111/
j.1532-7795.2011.00753.x.
Wang, M. T., & Eccles, J. S. (2012b). Social support matters: longitudinal effects of
social support on three dimensions of school engagement from middle to high
school. Child Development, 83, 877e895. http://dx.doi.org/10.1111/j.1467-
8624.2012.01745.x.
Wang, M. T., & Holcombe, R. (2010). Adolescents perceptions of school environ-
ment, engagement, and academic achievement in middle school. American
Educational Research Journal, 47, 633e662. http://dx.doi.org/10.3102/
0002831209361209.
Wang, M. T., & Huguley, J. P. (2012). Parental racial socialization as a moderator of
the effects of racial discrimination on educational success among African
American adolescents. Child Development, 83, 1716e1731. http://dx.doi.org/
10.1111/j.1467-8624.2012.01808.x.
Wang, M. T., & Peck, S. (2013). Adolescent educational success and mental health
vary across school engagement proles. Developmental Psychology.
 M.-T. Wang, J.S. Eccles / Learning and Instruction 28 (2013) 12e23
Wang, M. T., Willett, J. B., & Eccles, J. S. (2011). The assessment of school engage-
ment: examining dimensionality and measurement invariance across gender
and race/ethnicity. Journal of School Psychology, 49, 465e480. http://dx.doi.org/
10.1016/j.jsp.2011.04.001.
Wentzel, K. R. (2003). Sociometric status and adjustment in middle school: a lon-
gitudinal study. Journal of Early Adolescence, 23, 5e28. http://dx.doi.org/10.1177/
0272431602239128.
Wigeld, A., Eccles, J. S., Davis-Kean, P., Roeser, R., & Scheifele, U. (2006). Motivation
to succeed. In N. Eisenberg (Series Ed.) & W. Damon (Vol. Ed.) (6th ed.).,
23
Handbook of child psychology: Vol. 3. Social, emotional, and personality develop-
ment (pp. 933e1002) New York: Wiley.
Wigeld, A., Eccles, J. S., & Pintrich, P. R. (1996). Development
between the ages of eleven and twenty-ve. In D. C. Berliner, & R. C. Calfee
(Eds.), The handbook of educational psychology). New York: Macmillan
Publishing.
Zimmerman, B. J. (1990). Self-regulated learning and academic achievement:
an overview. Educational Psychologist, 21, 3e17. http://dx.doi.org/10.1207/
s15326985ep2501_2.